pPage

Teach Yourself Visual C++ 6 in 21 Days

---

- 21 -
Adding Web Browsing Functionality to Your Applications

• The Internet Explorer ActiveX Model
• The CHtmlView Class
  • Navigating the Web
  • Controlling the Browser
  • Getting the Browser Status
• Building a Web-Browsing Application
  • Creating the Application Shell
  • Adding Navigation Functionality
• Summary
• Q&A
• Workshop
  • Quiz
  • Exercises
• In Review

---

When Microsoft made the decision a few years ago to make all its applications Internet-enabled, it wasn't just talking about making Word read and write HTML pages. It wanted to make the Internet an integrated part of every application, in some way or another. Well, when it comes to development tools, making the editor double as an email client isn't really a practical integration. However, making it easy for the users of development tools to build Internet-enabled applications is a very practical feature. And this is exactly what Microsoft did.

One of the capabilities that Microsoft made available to its application development tools is using Internet Explorer as an integrated part of any application. This means that you can include Internet Explorer, and all its associated components, in your own applications. The possibilities extend far beyond providing your users Web browsing capability; your applications can also house, and interact with, Java applets. You can provide your users with not one, but two macro languages, VBScript and JScript (Microsoft's version of JavaScript).

Today, you will learn

• How the Internet Explorer ActiveX Object Model enables you to integrate all the components into your applications.

• How the CHtmlView view class encapsulates most of the Internet Explorer functionality in a ready-made class.

• How to build a simple Web browser using the CHtmlView class and Internet Explorer.

The Internet Explorer ActiveX Model

When Microsoft came up with the idea of integrating ActiveX with its Web browser, Internet Explorer, it realized that it would need to reengineer Internet Explorer to support the use of ActiveX controls. Well, the developers looked at what they would need to do, and what was possible, and decided to make Internet Explorer a lot more than just a Web browser.

The first thing that Microsoft did was separate the Web browser from the ActiveX objects that perform all the work. As a result, it ended up with the Internet Explorer application, which is little more than an ActiveX document container, and the Internet Explorer HTML viewer control, which ran as an ActiveX document server inside the application. This meant that the Internet Explorer application could host more than just Web pages; it could also be used to host Word documents, Excel spreadsheets, PowerPoint presentations, and any other ActiveX document that had an ActiveX document server installed on the same computer, as shown in Figure 21.1.

Within the HTML viewer component, Microsoft added the capability to host other controls, including scripting engines and ActiveX controls, as shown in Figure 21.2. This gave Microsoft the flexibility to add more scripting languages to Internet Explorer as they were requested and created. This also enabled Internet Explorer to host any ActiveX controls that developers might want to create.

In designing Internet Explorer this way, Microsoft not only gave itself a lot of flexibility for future expansion of the functionality supported by Internet Explorer, but it also made the entire workings of Internet Explorer available to any developer that wants to take advantage of it and integrate Internet Explorer into his or her applications.

FIGURE 21.1. The Internet Explorer ActiveX document model.

FIGURE 21.2. The Internet Explorer HTML viewer ActiveX object model.

The CHtmlView Class

To make it easy to incorporate the Internet Explorer HTML viewer into Visual C++ applications, Microsoft wrapped it in the CHtmlView class. This class can be used as the base class for the view class in your Single Document Interface (SDI) or Multiple Document Interface (MDI) applications. You can easily create applications that have built-in Web browsing capabilities.

Navigating the Web

Several functions available in the CHtmlView class cover navigating the Web. There are functions for returning the browser to the starting page of the user or for taking the user to an Internet search page. There are also functions for taking the user to the previous or next page or even to a remote Web page. All these functions are members of the CHtmlView class and thus are member functions of your application view class (when using the CHtmlView class as the base class for your view class).

The navigation functions for the CHtmlView class are listed in Table 21.1.

TABLE 21.1. CHtmlView NAVIGATION FUNCTIONS.

Function Definition	Description
GoBack()	Takes the user to the previous Web page.
GoForward()	Takes the user to the next Web page. (This assumes that the user has backed up from at least one Web page.)
GoHome()	Takes the user to the start page for Internet Explorer.
GoSearch()	Takes the user to an Internet search page.
Navigate(LPCTSTR URL)	Takes the user to the Web page specified in the URL variable.

The first four functions do not take any arguments and perform the exact same function as their toolbar equivalents in Internet Explorer. The last function does take arguments; the only required argument is the URL of the Web page to display.

Controlling the Browser

Along with the functions for navigating around the Web, you use some functions for controlling the browser. Two of these functions are Refresh(), which makes the HTML viewer control reload the current Web page, and Stop(), which halts a download in progress. As with most of the navigation functions, these functions do not take any arguments and work just like their equivalent toolbar buttons in Internet Explorer.

Getting the Browser Status

Another category of functions that are available in the CHtmlView class is informational in nature. You can use these functions to get information about the current state of the browser. For instance, if you want to get the current Web page in the browser, you can call GetLocationURL(), which returns a CString containing the URL. If you want to determine if the browser is busy with a download, you can call GetBusy(), which returns a boolean value specifying whether the browser is busy.

Many more functions are available in the CHtmlView class, and some of them only work on Internet Explorer itself, not on the browser control.

Building a Web-Browsing Application

For an example of how you can integrate the Internet Explorer Web browser component into your own applications, you will build a simple Web browser application. You will create an SDI application using the CHtmlView class as the base for your own view class. You'll add a menu with functions for the back and forward navigation options. You'll also add a dialog for getting from the user a URL that you will use to navigate the browser to the specified Web page.

Creating the Application Shell

To create a Web browser application, you can create a standard SDI or MDI application shell. The only other thing that you need to ensure is that Internet Explorer is installed on the computer where your application will run. For your development computer, this is not a problem because the Visual C++ installation probably required you to install the latest version of Internet Explorer. On any computers where you run your application, however, you might need to make sure that Internet Explorer is installed or install it yourself.

To create the shell of the application that you will build today, start a new project using the MFC AppWizard to create the application shell. Give the project a suitable name, such as WebBrowse, and click the OK button to start the AppWizard.

In the AppWizard, you can just as easily create an MDI Web browsing application as you can create an SDI application. For the purposes of the sample application that you are building today, go ahead and specify that the application is a Single document (SDI) application. You can accept the default settings for the rest of the AppWizard; for this example, however, choose the Internet Explorer ReBars for the toolbar appearance on the fourth AppWizard step.

Finally, on the sixth step, specify the CHtmlView class as the base class for your view class. This causes your application to be created using the Internet Explorer Web browser control as the main application view.

Once you finish generating the shell for your application, if you compile and run it while connected to the Internet, you'll find that you already have a working Web browser, as shown in Figure 21.3. However, you do not have the ability to specify where your browser will take you, other than clicking links in the Web pages displayed.

FIGURE 21.3. The initial Web browsing application.

Adding Navigation Functionality

Now that you've got a working Web browser, it would be nice if you could control where it takes you. What you need to add is an edit control where the user can enter a URL. Looking at the toolbar of the running application, you notice there's a place to put this control.

Specifying a URL

You probably noticed when you ran your application that the second toolbar had some static text telling you to lay out the dialog bar. The dialog bar is different from what you have worked with before. It is a toolbar with dialog controls on it. You even design the bar in the dialog layout designer. When you look for this dialog bar in the resource tab, you won't find it in the toolbar folder; it's in the dialogs folder.

If you open the dialogs folder and double-click the IDR_MAINFRAME dialog to open it in the dialog designer, you'll see that it's the second toolbar in your application. You can place edit boxes, buttons, combo boxes, and check boxes on this toolbar. You can place any control that you can use on a dialog window on this toolbar.

For your Web browser, modify the static text control already on the dialog bar and add an edit box, as shown in Figure 21.4. Specify an ID for the edit box; for this example, use the ID IDC_EADDRESS.

Before you open the Class Wizard to begin adding variables and event functions to the dialog bar, be aware that the dialog bar will automatically send its events to the main frame class in your application. When you open the Class Wizard, it assumes that you need to associate the dialog bar with a class and prompts you to create a new class. This association is not necessary because you can map all its events through the frame and from there feed them to the view or document classes.

FIGURE 21.4. The dialog bar layout.

For this example, you don't even need to use the Class Wizard to add any event handlers to the dialog bar. You need to trigger an action when the user finishes entering a URL into the edit box. The closest event available to you through the Class Wizard is the EN_CHANGED event, which will trigger for each letter the user types. What you need is an event that will trigger when the user presses the Enter key. Fortunately, when the user types in the edit box on the dialog bar and presses the Enter key, the IDOK command ID is sent to the frame class. What you can do is add a command handler in the message map to call a function on the IDOK command.

In your command handler, you need to get the window text from the edit box on the dialog bar. You can pass this string to the Navigate function in the view class, making the browser go to the page specified by the user.

To add this functionality to your application, add a new member function to the CMainFrame class. Specify the function type as void, the function declaration as OnNewAddress, and the access as public. Edit the new function, adding the code in Listing 21.1.

LISTING 21.1. THE CMainFrame OnNewAddress FUNCTION.

1: void CMainFrame::OnNewAddress() 2: { 3:     CString sAddress; 4: 5:     // Get the new URL 6:     m_wndDlgBar.GetDlgItem(IDC_EADDRESS)->GetWindowText(sAddress); 7:     // Navigate to the new URL 8:     ((CWebBrowseView*)GetActiveView())->Navigate(sAddress); 9: }

In this function, line 6 got the text in the edit box using the GetWindowText function, placing the text into the m_sAddress variable. The dialog bar was declared in the CMainFrame class as the m_wndDlgBar variable, so you were able to use the GetDlgItem function on the dialog bar variable to get a pointer to the edit box.

In line 8, you cast the return pointer from the GetActiveView function as a pointer to the CWebBrowseView class. This allowed you to call the Navigate function on the view class, passing it the URL that was entered into the edit box.

Now that you are able to take the URL that the user entered and tell the browser component to go to that Web page, how do you trigger this function? You have to add the message-map entry by hand because this is one that the Class Wizard isn't able to add. In the message map, after the closing marker of the AFX_MSG_MAP (the section maintained by the Class Wizard), add the ON_COMMAND macro, specifying the IDOK command and your new function as the handler to be called, as in Listing 21.2. You can also add this entry before the Class Wizard section as long as it's on either side and not inside the section maintained by the Class Wizard.

LISTING 21.2. THE CMainFrame MESSAGE MAP.

1: BEGIN_MESSAGE_MAP(CMainFrame, CFrameWnd) 2:     //{{AFX_MSG_MAP(CMainFrame) 3:         // NOTE - the ClassWizard will add and remove mapping macros             Âhere. 4:         //    DO NOT EDIT what you see in these blocks of generated             Âcode ! 5:     ON_WM_CREATE() 6:     //}}AFX_MSG_MAP 7:     ON_COMMAND(IDOK, OnNewAddress) 8: END_MESSAGE_MAP()

If you compile and run your application, you can enter a URL into the edit box on the toolbar and press the Enter key, and your application should browse to the Web page you specified, as in Figure 21.5.

FIGURE 21.5. Browsing to a specified URL.

Displaying the Current URL

When surfing the Web, you often follow links on Web pages that take you to other Web sites. When you do this, you wouldn't know what Web site you accessed if your browser didn't place the URL into the address box, indicating the site where you are and providing the opportunity to copy or modify the URL to find another page on the same site.

Getting the current URL from the browser is a simple matter of calling the GetLocationURL function and passing the result to the dialog bar. The problem is when to get the URL. It turns out that some event functions in the CHtmlView class can be overridden in your class. These functions will be triggered on various events that are triggered by the browser control. There are event functions for starting the navigation, beginning a download, monitoring a download's progress, and, most important for our needs, indicating a download has finished. None of these event handler functions can be added to your view class through the Class Wizard. You have to add them all.

To add the download-complete event handler to your application, add a new member function to the view class of your application. Specify the function type as void, the function declaration as OnDocumentComplete(LPCTSTR lpszUrl), and the access as public. Edit the function, adding the code in Listing 21.3.

LISTING 21.3. THE CWEBBROWSEVIEW ONDOCUMENTCOMPLETE FUNCTION.

1: void CWebBrowseView::OnDocumentComplete(LPCTSTR lpszUrl) 2: { 3:     // Pass the new URL to the address bar 4:     ((CMainFrame*)GetParentFrame())->SetAddress(lpszUrl); 5: }

You'll notice in this function that you didn't need to call the GetLocationURL function after all. The URL that is downloaded is passed as an argument to this function. This allows you to pass the URL along to the frame, where you'll add another function to populate the edit box on the dialog bar with the URL.

To add the function to populate the dialog bar with the new URL, add a member function to the main frame class, CMainFrame. Specify the function type as void, the function declaration as SetAddress(LPCTSTR lpszURL), and the access as public. Edit the function, adding the code in Listing 21.4.

LISTING 21.4. THE CMainFrame SetAddress FUNCTION.

1: void CMainFrame::SetAddress(LPCTSTR lpszURL) 2: { 3:     // Set the new URL in the address edit control 4:     m_wndDlgBar.GetDlgItem(IDC_EADDRESS)->SetWindowText(lpszURL); 5: }

In this function, you took the opposite path from the one you used to get the text from the edit box. You used the SetWindowText to change the text in the edit box to the URL that you are passing in. When you run your application, you should be able to see the URL address on the dialog bar change to reflect the Web page that you are viewing.

Back and Forth

Now that you can enter a URL into the dialog bar and have your application go to that Web site, and you can see the address of any Web sites that you view, it'd be nice if you could back up from where you might have gone. This is a simple matter of calling the GoBack and GoForward functions on the view class in your application. You can call these functions from menu entries, which also allows you to attach toolbar buttons to perform the same calls.

To add this functionality, open the main menu in the Menu Designer. You can delete the Edit menu from the bar, and all of the entries below it, because they are of no use in the application that you are building today. Grab the blank menu entry on the bar, and drag it to the left of the Help menu. Open the properties dialog for this menu entry and give it a caption of &Go. This is the menu where all navigation functions will be located.

To provide the back-and-forth functionality, you need to add two menu entries, one for the GoBack function and one for the GoForward function. Specify the properties for these two menu entries as shown in Table 21.2.

TABLE 21.2. MENU PROPERTY SETTINGS.

Object	Property	Setting
Menu Entry	ID	IDM_GO_BACK
	Caption	&Back\tCtrl + B
	Prompt	Back to the previous page\nBack
Menu Entry	ID	IDM_GO_NEXT
	Caption	&Next\tCtrl + N
	Prompt	Go forward to the next page\nNext

Once you add the menu entries, you can use the Class Wizard to add functions to the view class on both of these menu events. For the IDM_GO_BACK menu ID, add an event function on the COMMAND event message. Edit the function, adding the code in Listing 21.5.

LISTING 21.5. THE CWebBrowseView OnGoBack FUNCTION.

1: void CWebBrowseView::OnGoBack() 2: { 3:     // TODO: Add your command handler code here 4: 5:     // Go to the previous page 6:     GoBack(); 7: }

Open the Class Wizard again, and add an event-handler function for the IDM_GO_NEXT object ID on the COMMAND event message. Edit this function with the code in Listing 21.6.

LISTING 21.6. THE CWebBrowseView OnGoNext FUNCTION.

1: void CWebBrowseView::OnGoNext() 2: { 3:     // TODO: Add your command handler code here 4: 5:     // Go to the next page 6:     GoForward(); 7: }

Now you can run your application and use the menus to back up to the previous Web pages from wherever you surfed to and then trace your steps forward again. However, it's somewhat difficult using the menus, so what you need to do is add an accelerator for each of these menu entries.

If you open the accelerator table in the resources tree, you see a bunch of accelerators tied to menu IDs. Each of these accelerators consist of an ID and a key combination. If you right-click anywhere in the accelerator table, you see the option of adding a new accelerator to the table. Choosing this option presents you a dialog to enter the accelerator information. First, you need to specify the menu ID that the accelerator will be tied to. (As with toolbar buttons, accelerators are tied to menu entries.) Below that, you can enter the key that will trigger the accelerator, or you can select a key from the drop-down list.

On the right side of the dialog, you can select the modifiers for the key. Modifiers are the other keys that must be pressed in combination with the key that you've already specified for the accelerator to be triggered. Once you've entered all the necessary information for the accelerator, close the dialog and the information you specified is added to the table.

---

TIP: It's recommended that you use either the Ctrl or Alt key as one of the modifier keys on all accelerators using standard keys. If you don't use one of these two keys as part of the accelerator, your application might get confused about when the user is typing information into your application and when the user is triggering an accelerator.

---

To add accelerators to the back and forward menus in your application, delete the accelerator for the ID_FILE_OPEN menu ID because you won't use it in this application. Add a new accelerator and specify the ID as IDM_GO_BACK and the key as B and select the Ctrl modifier. Add a second accelerator, specifying the ID as IDM_GO_NEXT and the key as N and select the Ctrl modifier. When you run your application, you can use the Ctrl+B key combination to back up to the previous page and the Ctrl+N key combination to go forward.

To really make your application work like most available Web browsers, you would also add toolbar buttons for these two menu entries with arrows pointing to the left for back and to the right for forward.

Controlling the Browser

Often when browsing, you come across a Web page that you don't want to wait to download. You'll want to stop the transfer part-way through. Maybe you entered the wrong URL or maybe the download is taking too long. It doesn't matter why you want to stop the download; it's enough that you want to stop it. This is why the CHtmlView class has the Stop function. It cancels the download currently in progress. To add this functionality to your application, add a new menu entry to the View menu in the Menu Designer. Specify the menu entry properties in Table 21.3.

TABLE 21.3. MENU PROPERTY SETTINGS.

Object	Property	Setting
Menu Entry	ID	IDM_VIEW_STOP
	Caption	Sto&p
	Prompt	Stop the current transfer\nStop

Using the Class Wizard, add an event-handler function to the view class for this menu ID on the COMMAND event message. Edit the function with the code in Listing 21.7.

LISTING 21.7. THE CWebBrowseView OnViewStop FUNCTION.

1: void CWebBrowseView::OnViewStop() 2: { 3:     // TODO: Add your command handler code here 4: 5:     // Stop the current download 6:     Stop(); 7: }

If you run your application, you can use this menu entry to stop any download of a Web page that you don't want to wait on. It would be more convenient if you added a toolbar button for this menu ID.

Another control function that most browsers have is the capability to reload the current Web page. This function is handy for Web pages that contain dynamic elements that change each time the page is downloaded. It's also helpful for Web pages that your browser may have in its cache so that it doesn't retrieve the newest version of the page. It's necessary to be able to force the browser to reload the page and not just display the cached version (especially if it's a Web page that you are in the process of creating). The browser component has the capability built in with the Refresh function. One call to this function means the current page is reloaded.

You can add this functionality to your application by adding another menu entry to the View menu. Specify the properties for the new menu entry using the settings in Table 21.4. You can add a separator bar between the two View menu entries that were originally there, and the two new entries, to make your menu look like the one in Figure 21.6.

TABLE 21.4. MENU PROPERTY SETTINGS.

Object	Property	Setting
Menu Entry	ID	IDM_VIEW_REFRESH
	Caption	&Refresh
	Prompt	Refresh the current page\nRefresh

FIGURE 21.6. The modified View menu.

Once you add the menu entry, use the Class Wizard to add an event-handler function to the view class for the COMMAND event message for this menu entry. Edit the function, adding the code in Listing 21.8.

LISTING 21.8. THE CWebBrowseView OnViewRefresh FUNCTION.

1: void CWebBrowseView::OnViewRefresh() 2: { 3:     // TODO: Add your command handler code here 4: 5:     // Reload the current page 6:     Refresh(); 7: }

Now you can test this functionality by finding a Web site that returns a different page each time that you refresh the browser, such as the Random Monty Python Skit Server in Figure 21.7. As with the rest of the menu functions that you added to this application, this one should also be added to the toolbar.

FIGURE 21.7. The Refresh function will perform a new download.

Summary

Today you learn how Microsoft designed its Internet Explorer Web browser as a series of ActiveX components that could be used in other applications. You saw how Microsoft encapsulated the browser into the CHtmlView class, which can be used in SDI and MDI applications to provide Web browsing functionality to almost any application. You learned how you can use this view class to build a Web browser. You saw how you could use the dialog bar to place controls on a toolbar and how the events for these controls can be handled in the frame class for the application. Finally, you learned how to add menus to your application to call the various functions of the Web browser to provide a complete surfing experience.

Q&A

Q Why is Print Preview not included on the default menus when I choose CHtmlView as the base class for my view class?

A The printing for the CHtmlView class is performed by the browser, not the view class. You don't have print preview because the browser doesn't support it.

Q How can I get the HTML source code from the browser so that I can see or edit it?

A The CHtmlView class has a member function, GetHtmlDocument, that returns a pointer to the IDispatch interface of the document object in the browser. You have to use the Invoke IDispatch function to call the functions in the document object to retrieve the HTML source code. For documentation on how to do this, check out the Microsoft Developer Network CD that you received with your copy of Visual C++ 6.

Workshop

The Workshop provides quiz questions to help you solidify your understanding of the material covered and exercises to provide you with experience in using what you've learned. The answers to the quiz questions and exercises are provided in Appendix B, "Answers."

Quiz

1. What does the CHtmlView class encapsulate for use in Visual C++ applications?

2. How can you get the URL for the current Web page from the CHtmlView class?

3. What command is triggered for the frame class when the user presses the Enter key in the edit box on the dialog bar?

4. What functions can you call to navigate the browser to the previous and the next Web pages?

5. How can you stop a download in progress?

Exercises

1. Add the GoSearch function to the menu and toolbar.

2. Add the GoHome function to the menu and toolbar.

Disable the Stop toolbar button and menu entry when the application is not downloading a Web page.

In Review

You made it! You now have the skills to tackle most of the Windows programming tasks in Visual C++, with the exception of those that require specialized knowledge and understanding. Although there's still plenty for you to learn, you have covered the vast majority of the topics in a very short amount of time. From here you will probably want to delve into one or two more specialized areas of Windows programming--the type that an entire book can be written on--because at this point, you've got just about all the generalized skills down.

Just in case you don't have them all down, it's not a bad idea to take some time once again to come up with some of your own applications where you can apply the things you've learned. This will help pinpoint any areas that you might need to go back and review before jumping into any more advanced topics. Let's take a quick look back at what you should have learned during the past week, just to make sure.

You started off the week by learning about Microsoft's latest database application development technology, ActiveX Data Objects (ADO). You learned how ADO sits on top of the OLE DB technology, simplifying database access considerably. You learned how you can easily build a database application by using ADO as an ActiveX control and connecting it to some ADO-compatible, data-bound ActiveX controls. You also learned how you can import the DLL and dive into the code, exerting complete control over how your application deals with the database. You learned how you can read and write each of the fields in a record in the record set by converting it from and to the variant data type. You also learned how you can design your own custom record class and can bind it to the record set object, providing you with a much more convenient way to access the individual field values in the database. If you are having problems with any of this, you might want to review Day 15 once more.

You learned three different ways of sharing the functionality modules that you develop with other programmers, without having to share your code. First you learned how to build your modules into library files that can be linked into applications by other Visual C++ developers. You saw how with this approach, if any change is made to the library module, then all the applications that it's used in have to be rebuilt. You learned how you can create these modules without making any special changes to the way you write your code. If you don't remember how you did all of this, you can go back to Day 16 to review it.

The second way that you learned to share your functionality with other programmers was to build DLLs. You learned how you can create two different types of DLLs: one that can be used only by other Visual C++ applications and one that can be used by any other application, regardless of what programming language was used to build it. You saw how you can create a DLL for use by other Visual C++ programmers without having to make any real changes to the way you design or code your modules. You also learned how you need to make dramatic changes to how your module is used and interacted with when creating DLLs that can be used by all programming languages. You learned how to provide straight function calls as an interface for use by other applications, with all necessary information to be passed in as parameters to the functions. You also learned how to build a definition file, with the names of all functions to be exported included in it. If you need any reminders of how you can do any of this, you'll want to look over Day 17 again.

The third and final way that you learned to share the functionality of your modules without sharing the code was to package it as an ActiveX control. You learned how to use the Control and Class Wizards to add in all the properties, methods, and events that you want to have in your control. You learned how to read and write the properties in your control. You saw how there are two different ways that the properties in your control can be implemented, and you learned how to determine which type is appropriate for each of your control's properties. You learned how you can raise events in the container application from your control by firing the event in your code. Along with all of this, you learned how you can use the ActiveX Control Test Container utility to test your control, calling all its methods and triggering all the events that it's capturing. You saw how you can monitor the events that your control is firing in the containing application to make sure that they are being fired as and when they should. If you need any reminders of how all this works, you can look back at Day 19 for a refresher.

An important thing that you learned was how you can enable your applications to perform more than one task at a time. This is an important piece of functionality, and more applications are requiring this capability every day. Not only did you learn how to make your applications perform multiple tasks at once, but you also learned two different ways to do so. First, you learned about the OnIdle function and how you can hook into this function to trigger your own functionality to be run when the application is sitting idle. You also learned about the shortcomings of using this approach to adding a second task to your application and how it can prevent your application from responding to the user. You need to slice the background task into little pieces that can be done quickly, which requires you to develop some elaborate way of keeping track of where the task is and where it needs to pick back up when the application is idle again.

The second way that you learned to give your applications a second or third task to do is by spinning off separate threads, which run completely independent of the main user-interface thread. You learned how to create a callback function that controls the top level of execution for the threads and how you can start and stop the thread as necessary. You also saw how these independent threads are completely independent from the rest of the application and how they'll continue to run, even when the rest of the application is also busy. If you feel the need to look at all this a second time, you might want to read Day 18 again.

Another area of growing importance that you learned about was how to build Internet applications using the Winsock interface classes. You learned how you can build one application that connects to another over a network and sends messages back and forth. You learned that, just like with a telephone, for one application to connect to another, the second application has to be listening for the connection. You saw how easy it is to send messages and to be notified when a message has arrived after the connection between the two applications has been made. If you need to review some of this, you might want to look back at Day 20.

Finally, you learned how you can incorporate the Microsoft Internet Explorer Web browser into your application without any effort whatsoever. You learned how you can control the browser by specifying the URL that it should load and display for the user and how you can display informational messages to the user to show what the browser is doing and when it's busy. If you need to look back at this to refresh your memory, you can go back to Day 21.

That's it. You're done. You've covered a lot of ground and learned some advanced topics, especially over this last week. Now it's time to put this book down and get busy programming, building your own applications using what you've learned. Good luck. If you find that you need a little help or advice, a good place to turn is the Microsoft newsgroups on the Internet. They are full of people who are both knowledgeable and helpful.

---

© Copyright, Macmillan Computer Publishing. All rights reserved.